ifLib is generating noise and fallout, including kernel panics. It is not so broken that it has to be reverted. Scott will go back and take a look at the changes as well.
Ryan is trying to inline the mbuf allocator. The problem is that this prevents using the uma constructors. Right now he's wrapped the uma_zalloc in an inline function. This is all still a WiP.
gnn has pushed a TCP test using dummynet into the netperf repo
With 10% loss on a back to back 10G link things get very bad but we're not sure if it's worse that the theoretically limit.
Jonathan has sent out an RA test but it needs to be more widely shared.
Michael pushed up D6452 to change the tun interface MTU and would like more review.
Hiren has shared a “strange pcap” from 10.2 on the transport@ list and would like others to look it over. One issue is that there seems to be a an artificial limit on the reassembly queue.
Hiren asks about tools to rewrite the IP addresses. gnn will pass some along.